Project Summary The human body is composed of a large number of cell types. These types are generally quite stable in the sense that cells of one type, once established, typically do not switch to other cell types. A major goal for regenerative medicine and our understanding of cell type in general is to discover if and how we can force cells to switch from one type to another. Recent results over roughly the last decade have shown that it is in principle possible to convert cells from one type by forcing them to turn on small sets of genes. However, in the vast majority of cases, we have no idea what these sets of genes are out of the many thousands of potential ones, and so our understanding has largely been dictated by picking candidates based on prior knowledge. Furthermore, even when these sets of genes are identified, the efficiency of interconversion of cell type is very low, with only a small percentage of source cells converting to the target type. Our proposed research tackles both of these challenges using a combination of new concepts of cell identity and new technology for tracing individual cells back in time. For cell identity, the approach most common in the field is to use profiles of which genes are on or off in any particular cell type to determine lineage-specific factors. However, while these genes are lineage-specific, they may not be lineage-determining in the sense that they may not drive a cell to a particular type per se. We have developed a methodology we call PerturbID that uses a series of systematic perturbations to identify specific genes that turn on and off in response, which we have shown have the capacity to interconvert cells. We propose to use PerturbID to identify and validate candidates for cell type transformation across a range of cell types, ultimately arriving at a set of general principles for cellular reprogramming. This will have applications for regenerative medicine as well as across biology as a whole. The other major problem, of efficiency, has remained mysterious because nobody currently knows why some cells are capable of reprogramming while the vast majority are not. The challenge is the lack of tools for retrospective profiling of cells: how do we rewind time to profile the cells that will ultimately adopt a different fate? We have developed tools for performing this retrospective profiling. We will apply this ?time machine? methodology to the problem of inefficient reprogramming to determine the unique signature of cells primed for cell type conversion, and will perform genetic screens to isolate pathways capable of manipulating this frequency. Together, our work will transform our concepts of cell type and will have enormous practical implications for its application in regenerative medicine.